Ambient AI as a Tool for Returning Time to Care

Ambient AI scribes offer the ability to return clinician time to the patient encounter by taking on a meaningful share of the documentation work that pulled clinicians away from patients in the first place. The question now is whether that returned time benefits the clinicians who need it the most and whether it comes at any cost to how patients experience their care. The authors report findings from an enterprise-wide ambient AI scribe deployment across the ambulatory practices of a large integrated health system in Eastern Pennsylvania and Southern New Jersey, conducted 1 year following an organizational merger. Among 5234 provisioned clinicians, the authors analyzed 283,028 ambulatory encounters that occurred over a 6-month period (July 1-December 31, 2025) with matched patient experience surveys, with results reported across two implementation cohorts reflecting distinct legacy electronic health record instances and prior scribe program histories. Documentation efficiency was assessed in a 637-clinician analytic cohort, meeting threshold utilization criteria of 50 or more ambient encounters and at least 10% utilization. Among 1802 active clinicians, the platform delivered an 84% mean effort reduction in completion of history of present illness and assessment and plan note sections, with an average of 14.9 minutes of recording per encounter. Epic Signal data showed after-hours documentation ("pajama time") dropping 8.9% and 16.6% across the two cohorts, time in notes per appointment falling 10.5% and 17.3%, and time to chart closure decreasing 25.3% and 39.3%. Top-quartile improvers gained back nearly 17 minutes of pajama time per scheduled day and reduced time to chart closure by more than 36 hours. Benefits concentrated among the most documentation-burdened clinicians: Advanced practice clinicians experienced approximately 2.5 times the daily documentation reduction observed among physicians (-9.9 vs. -3.8 min/day, P<0.001), consistent with a floor-raising effect (largest gains among the most documentation-burdened clinicians) rather than uniform effect. Dose-response relationships were statistically significant for time in notes per day (P<0.001), time to chart closure (P=0.002), and same-day closure (P<0.001). Patient experience showed no evidence of harm. Measured with two different experience survey instruments across the cohorts, ambient AI encounters scored equivalent to or higher than non-ambient AI encounters across all domains. Within-provider analyses mitigated these differences and were largely nonsignificant, consistent with clinician self-selection. The integration and governance approach reported here - anchored by defined organizational purpose, executive sponsorship, electronic health record-native measurement, and voluntary clinician-centered adoption - offers transferable implementation principles for health systems. Most notably, ambient AI functioned as a floor-raiser, returning the most time to the clinicians most in need of it. This provides a workforce equity implication with direct relevance for how systems target and support ambient AI scribe deployment.

Authors

Institutions

Publication Details

Journal
NEJM Catalyst
Published
2026-11-18
DOI
https://doi.org/10.1056/cat.26.0269
Citations
1
Primary Topic
Electronic Health Records Systems
Type
article
Field-Weighted Citation Impact
2.64
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Ambient AI as a Tool for Returning Time to Care

Bracken Babula, Patricia C. Henwood, Baligh R. Yehia, Melanie McArthur et al.
1 citations
NEJM Catalyst
Electronic Health Records Systems
2.64
article

Ambient AI as a Tool for Returning Time to Care

Bracken Babula, Patricia C. Henwood, Baligh R. Yehia, Melanie McArthur, Jennifer Stephens
article en
1 citations

Abstract

Ambient AI scribes offer the ability to return clinician time to the patient encounter by taking on a meaningful share of the documentation work that pulled clinicians away from patients in the first place. The question now is whether that returned time benefits the clinicians who need it the most and whether it comes at any cost to how patients experience their care. The authors report findings from an enterprise-wide ambient AI scribe deployment across the ambulatory practices of a large integrated health system in Eastern Pennsylvania and Southern New Jersey, conducted 1 year following an organizational merger. Among 5234 provisioned clinicians, the authors analyzed 283,028 ambulatory encounters that occurred over a 6-month period (July 1-December 31, 2025) with matched patient experience surveys, with results reported across two implementation cohorts reflecting distinct legacy electronic health record instances and prior scribe program histories. Documentation efficiency was assessed in a 637-clinician analytic cohort, meeting threshold utilization criteria of 50 or more ambient encounters and at least 10% utilization. Among 1802 active clinicians, the platform delivered an 84% mean effort reduction in completion of history of present illness and assessment and plan note sections, with an average of 14.9 minutes of recording per encounter. Epic Signal data showed after-hours documentation ("pajama time") dropping 8.9% and 16.6% across the two cohorts, time in notes per appointment falling 10.5% and 17.3%, and time to chart closure decreasing 25.3% and 39.3%. Top-quartile improvers gained back nearly 17 minutes of pajama time per scheduled day and reduced time to chart closure by more than 36 hours. Benefits concentrated among the most documentation-burdened clinicians: Advanced practice clinicians experienced approximately 2.5 times the daily documentation reduction observed among physicians (-9.9 vs. -3.8 min/day, P<0.001), consistent with a floor-raising effect (largest gains among the most documentation-burdened clinicians) rather than uniform effect. Dose-response relationships were statistically significant for time in notes per day (P<0.001), time to chart closure (P=0.002), and same-day closure (P<0.001). Patient experience showed no evidence of harm. Measured with two different experience survey instruments across the cohorts, ambient AI encounters scored equivalent to or higher than non-ambient AI encounters across all domains. Within-provider analyses mitigated these differences and were largely nonsignificant, consistent with clinician self-selection. The integration and governance approach reported here - anchored by defined organizational purpose, executive sponsorship, electronic health record-native measurement, and voluntary clinician-centered adoption - offers transferable implementation principles for health systems. Most notably, ambient AI functioned as a floor-raiser, returning the most time to the clinicians most in need of it. This provides a workforce equity implication with direct relevance for how systems target and support ambient AI scribe deployment.

NEJM CatalystVol. 7(12)
Thomas Jefferson University (US), Thomas Jefferson University Hospital (US), Jefferson College (US)
Good health and well-being
Openalex Percentile: Top 15%
Electronic Health Records Systems
2.64
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon